TEER in Functional MR: Data Deluge, but No Easy Answers

At ESC, we learned the positive but nuanced results of RESHAPE-HF2, the third trial for transcatheter edge-to-edge repair (TEER) in patients with functional mitral regurgitation (MR). 
The New England Journal of Medicine published the main paper, and the Journal of the American College of Cardiology (JACC) published a subgroup analysis, a meta-analysis, and celebratory and skeptical editorials.
Despite this deluge of data and expert opinion, nuance predominates over clear answers when it comes to treating MR due to heart failure. 
The Background: MITRA-FR and COAPT
The MITRA-FR trial of TEER added to optimal medical therapy vs OMT alone delivered neutral results. The French authors reported similar rates of heart failure events, cardiovascular (CV) death and overall death in both groups. 
In contrast, the COAPT trial reported that adding TEER dramatically reduced heart failure events, CV death, and overall death. The effect size for the primary endpoint of CV death and hospitalization for heart failure was a massive 32% in absolute terms — a number needed to treat (NNT) of 3. 
Papers published in the cardiovascular imaging specialty versions of the European Heart Journal and JACC: Cardiovascular Imaging have attempted to explain how seemingly similar trials with similar patients delivered divergent results. The core idea holds that the MR in COAPT patients contributed much more to their clinical demise than did the MR in MITRA-FR participants. In the French trial, the left ventricular dilation was the bigger problem — not the valve. This argument leans heavily on precise echocardiographic characteristics. 
RESHAPE HF2 
RESHAPE HF2 enrolled 505 patients with secondary MR, reduced left ventricular ejection fraction, and symptomatic heart failure despite medical therapy, and randomly assigned them to have TEER or not. These were 70-year-old mostly male patients (80%) with a median left ventricular ejection fraction (LVEF) of 33%. 
The investigators chose three endpoints. The first was a composite of first or recurrent hospitalization for heart failure or CV death; the second was first or recurrent hospitalization for heart failure; and the third was change from baseline to 12 months in the score on the Kansas City Cardiomyopathy Questionnaire (KCCQ).
The baseline clinical characteristics in RESHAPE HF2 participants match up well with those in MITRA-FR and COAPT. The main difference was the general severity of the MR. RESHAPE-HF2 patients had lower mean B-type natriuretic peptide values, slightly better kidney function, less severe MR, and a lower mean effective regurgitant orifice area (EROA). The mean EROA was 0.25 cm2 compared with 0.31 cm2 in MITRA-FR and 0.40 cm2 in COAPT. A considerably lower fraction of patients in RESHAPE-HF2 had an EROA >0.40 cm2 compared to COAPT (9% vs. 41%).
Results
At 2 years, the rate of CV death or heart failure hospitalization was 37 events per 100 patient-years in the device arm vs 59 events per 100 patient years in the control group (rate ratio, 0.64; 95% CI, 0.48 – 0.85; P = .002).
First or recurrent hospitalization for heart failure was 27 per 100 patient years in the device arm vs 46.6 per 100 patient years in the control group (rate ratio, 0.59; 95% CI, 0.42 – 0.82; P = .002).
The mean difference in KCCQ score over 1 year was 10.9 points higher in the device arm. This too met statistical significance. 
More impressive risk reductions in the first primary endpoint were notable in subgroups with more severe heart failure; worse MR, higher EROA, and especially in patients who had been hospitalized for heart failure within a year before randomization. 
These were the results that made headlines. Let’s move now to the endpoints less emphasized on social and news media. 
Despite the 20% absolute risk reduction in heart failure events, the device did not produce a statistically significant reduction in all-cause mortality, cardiovascular death, or the rate of total hospitalizations for any cause. 
Six Reasons for Caution
While the results of RESHAPE-HF2 are positive, here are six reasons that the trial does not settle the tension between MITRA-FR and COAPT. 
First is the matter of external validity or generalizability. The trial took 8 years to recruit from 30 centers, each contributing only two patients per year. These were highly selected patients, and care should be taken in translating the results to clinical practice. 
Second is the open nature of the trial. The two drivers of benefit in RESHAPE-HF2 were heart failure hospitalizations and quality of life. Device trials are done at centers with doctors who love devices. These clinicians must decide to admit a patient with heart failure or give outpatient diuretics. Knowing treatment assignment could affect this decision. Patients also know if they underwent TEER, and those without the device could experience negative thoughts about having bad valvular disease and not getting the device. The only proper way to use quality-of-life measures as an endpoint is with a placebo-controlled (or sham) arm. 
Third is the lack of significant reductions in hard endpoints. CV death and all-cause death did not significantly differ. This could be due to low power, but you would think that a device that markedly reduces hospitalization for heart failure in older patients with substantial left ventricular systolic dysfunction should reduce CV and all-cause death. That it doesn’t just strengthens the argument that bias may have affected the hospitalization component of the endpoint. 
Fourth is the lack of effect on all-cause hospitalization. Similarly, you’d expect that a 20% absolute reduction in heart failure admissions would be enough to reduce all-cause hospitalization. That it did not suggests that either the effect size of this device in these patients is less robust than the primary outcome numbers reveal, or these patients have serious competing causes of hospitalization. The former seems most likely. 
A fifth concern is differential loss of data in the two groups. “Premature termination of the trial” occurred in 41 patients in the control arm vs 13 in the device arm. This supports the idea that patients (and clinicians) may have had some degree of bias in treatment assignment. This difference isn’t enough to negate the results, but it should make us cautious. 
A final concern comes from discussions I have heard (online) about using these results to expand TEER to patients with less severe MR. This is a bad idea. In addition to the five reasons for caution, multiple subgroup analyses (recent hospitalization, severity of MR, EROA size) all strongly suggest that TEER exerts most of its effect in patients who have the most severe forms of secondary MR. 
Final Conclusion
These three trials and my last few years of taking care of patients who have TEER convinces me that functional MR is a highly diverse condition. 
It’s hard to study diverse conditions. Each patient is so different that average results in a trial are hard to apply to an individual patient. 
The inherent diversity of this condition also makes me doubt that more trials in this space will make it any easier to identify those few patients likely to benefit from TEER. 
I would propose that the skill required to sort out the patient most likely to benefit from TEER may be greater than the skill needed to place the clips on the valve leaflets.
John Mandrola practices cardiac electrophysiology in Louisville, Kentucky, and is a writer and podcaster for Medscape. He espouses a conservative approach to medical practice. He participates in clinical research and writes often about the state of medical evidence.